Any one of a number of sources Navitoclax concentration of glutamate might activate presynaptic NMDARs. However, because the incidence of large transients was reminiscent of the stochastic pattern of transmitter release, we decided to block AP-evoked glutamate release and assess whether this changed the probability of observing large Ca2+ transients. Blocking neurotransmission by application of bafilomycin A1 (Figures 8Aii and 8Aiii) significantly reduced the probability of observing a large Ca2+ event (ACSF θ = 0.18 ± 0.067; Baf A1 θ = 0.008 ± 0.016; n =

5; Figure 8Aiv), suggesting that AP-evoked glutamate release is critical for the generation of large Ca2+ events and that presynaptic NMDARs are activated when an AP triggers this release. To guard against off-target

effects, we repeated the experiment by blocking neurotransmission with Botulinum toxin (BoTx) type C (500 nM). Dialysis of BoTx via micropipette into CA3 cells significantly reduced the probability of observing a large Ca2+ event in boutons (ACSF θ = 0.186 ± 0.067; BoTx θ = 0.008 ± 0.017; n = 4; (Figures 8Bi–8Biv), again consistent with the idea that transmitter release Ion Channel Ligand Library cell assay is necessary for the generation of large Ca2+ events. In light of these data, we propose a model that describes the way in which large Ca2+ transients arise from NMDAR activation (Figure 9). (1) AP invasion into the terminal depolarizes the membrane. The duration of a somatically recorded AP in hippocampal neurons is ∼2.5–3 ms (Qian and Saggau, 1999 and Gong et al., 2008). (2) Depolarization opens VDCCs, which elevates [Ca2+]i and produces Rebamipide a small Ca2+ transient in the bouton. The time taken to open VDCCs is ∼0.2 ms (Lee et al., 2000 and Randall and Tsien, 1995), and the time taken for diffusion of Ca2+ from VDCCs to synaptic vesicles is estimated to be ∼0.3 ms (Meinrenken

et al., 2002). (3) When release occurs, the time taken for exocytosis is ∼0.3 ms (Bruns and Jahn, 1995 and Meinrenken et al., 2002). Glutamate must then diffuse to the NMDARs. Previous studies report diffusion coefficient values in the range D   = 0.3–0.76 μ2/ms ( Savtchenko and Rusakov, 2004 and Ventriglia and Di Maio, 2000). For diffusion to occur across the width of a bouton (0.5–1 μm), we estimate the transition time (ttr  ) to be between 0.05 and 0.5 ms. This is determined by assuming that glutamate performs a random walk in which mean square deviation (MSD) is described by MSD=6⋅D⋅ttrMSD=6⋅D⋅ttr. By summing each of these parameters, we estimate that glutamate will arrive at the presynaptic NMDARs within 1.3 ms. (4) Because glutamate arrival occurs during the envelope of depolarization of the AP, relief of the Mg2+ block is concurrent with the arrival of glutamate. The kinetics of Mg2+ unblock are reported to be very fast, around 100–200 μs ( Jahr and Stevens, 1990 and Kampa et al., 2004), so Mg2+ relief looks unlikely to be rate limiting.